GNOME Bugzilla – Bug 543157
It is not always possible to distinguish on-screen text from Orca's "system" messages
Last modified: 2011-01-15 17:45:08 UTC
I would like to suggest that a fourth voice be added to orca. This voice would be configuredrin a manner similar to the Uppercase and hyperlink voices. This voice could be used to indicate additional information. Some suggestions are: 1. Discriminate information not part of application content, such as when switching between applications, pulling down menus,updates to progress bars, levels of indentation, etc. 2. Used to indicated selected text when navigating using flat review. 3. Indicate the misspelled word during spell check to avoid problems of homophones. 4. In text editors used for coding, it could indicate syntax highlighting. 5. It could indicate the nature of controls in forms, such as text, radio button, check box. I'll add more ideas as they occur to me. If the user does not wish to have this functionality, they configure the system voice to sound exactly like the default voice.
In bug 588910 comment #22, Jose asked: > Would be possible speak Misspelled Indicator using a different voice? This might be another job for the proposed "System Voice". Or not. But I don't want the idea to be lost, so I'm adding it here.
(In reply to comment #1) > In bug 588910 comment #22, Jose asked: > > > Would be possible speak Misspelled Indicator using a different voice? > > This might be another job for the proposed "System Voice". Or not. But I don't > want the idea to be lost, so I'm adding it here. > Hi, Joanie, The FRE for this is Bug 543160. I'm glad someone else is interested in this as well. :-) dave
Planning spam. Sorry!
*** Bug 591709 has been marked as a duplicate of this bug. ***
Created attachment 167897 [details] [review] Proposed fix for this bug and bug 412656 Due to upcoming freezes, I'm helping out some with the Andalucían bugs. What I've attached implements a "System" voice. In doing so, I also implemented the per-string ACSSes in the speech generators (i.e. this also fixes bug 412656). Wish I'd remembered 412656 when I'd started all this. :-) Anyhoo, Ale's done an initial review of these changes off-line. We'll solicit user feedback and testing now.
When navigating in a tree view, if the node is collapsed, the word collapsed is spoken using the default voice. This can be observed in pidgin and nautilus but seems work ok in thunderbird.
Created attachment 167905 [details] [review] Proposed fix for this bug and bug 412656 - Take 2 * Fixes the issue pointed out by José (thanks for testing!) * Moves getting the voice to the top of the generator methods in case we want to make a voice-based decision.
(In reply to comment #7) > Created an attachment (id=167905) [details] [review] > Proposed fix for this bug and bug 412656 - Take 2 > > * Fixes the issue pointed out by José (thanks for testing!) > Works fine, great job! Thanks.
Hi, Well I've been running with the patch for a while and Orca seems as stable as ever with it so I think it can go into master before the UI freeze. I do have some questions though. Should Orca say "blank" when you land on a blank line in the system voice? It doesn't right now and the word "blank" isn't on the screen. Also, when you receive a notification Orca says "notification" in the default voice. Is the word notification on screen? And when you activate a style change in OO writer, control + i for example, Orca says "italics on" in the default voice. Are these words on screen? I imagine it's just a toolbar item being highlighted. Paul
(In reply to comment #9) > Hi, > > Well I've been running with the patch for a while and Orca seems as stable as > ever with it so I think it can go into master before the UI freeze. Yay! Thanks Paul. > Should Orca say "blank" when you land on a blank line in the system voice? It > doesn't right now and the word "blank" isn't on the screen. True.... Probably not. > Also, when you receive a notification Orca says "notification" in the default > voice. Is the word notification on screen? Nope. Good catch. > And when you activate a style change in OO writer, control + i for example, > Orca says "italics on" in the default voice. Now you're just bein' picky. <huge, silly grin> Another good catch. Thanks! > Are these words on screen? I > imagine it's just a toolbar item being highlighted. You imagine correctly. Typically it's a letter 'i' (or 'b' or 'u') which appears to be italicized (or bolded or underlined). I'll take a look. Thanks again!!
So... On Paul's issues, I've got two solved and am still debugging the third. I'm confident that the changes attached/proposed thus far are sound and have been given thorough testing by me, the regression tests, and the user. This will not be the case with the changes for Paul's issues. Not only will they be largely untested, I'm less confident in their soundness. Therefore, my suggestion is that what's currently attached go into master (assuming Ale has no problems with it). What I'm working on for Paul's issues should not be committed until after 2.31.90 is out the door.
So far, Orca seems quite stable with the new system voice enhancement. I too have found a situation where the system voice is not being used when I think it should. I'm referring to Mozilla Firefox when reading down a page with the down arrow key. When Orca says things like "Link" or "Heading 1", "Heading 2", etc., these items are not being spoken in the new system voice; they speak in the default voice. Now when I go into other dialogs like the bookmarks dialog or pop up menus, the extra verbage like shortcut keys and the loading of pages and the "please wait" messages do speak in the system voice. So the biggest deal here is the details while reading page content. It can probably be addressed when Paul's findings are looked into; my observation doesn't require fixing before this week's release.
Thanks Steve. Confirmed. At one point I had that working. And it should be easy to get that re-working. <smile>
Comment on attachment 167905 [details] [review] Proposed fix for this bug and bug 412656 - Take 2 http://git.gnome.org/browse/orca/commit/?id=dbc8d2c388d721c30506bc5cdca97719b228e349 (I did catch and fix the one whitespace error. :-) ) I'll do String and UI announcements after any other commits we do today.
A couple more for you Joanie... When you press F6 to move to the read only pane in Pidgin, Orca says the "read only" bit in the default voice then the "text" bit in the system voice. Also, the "Orca user settings reloaded" message is spoken in the default voice. :)
And one from Attila: This debug.out shows how looks now the speech output generation with new committed system voice feature Dear Developers, After new system voice type feature is committed, little changed for example the heading level link text spokening intonation. Oldest Orca version send following type speech output before this feature with heading level links, little part with an oldest debug.out file: SPEECH OUTPUT: 'Kukásautóban a konténerben alvó férfi hivatkozás 1. címsor szint' voice=hyperlink After new System voice feature is committed, the speech output is following changed if generated speech result for a heading level link role: SPEECH OUTPUT: 'Kukásautóban a konténerben alvó férfi hivatkozás' voice=hyperlink SPEECH OUTPUT: '1. címsor szint ' voice=hyperlink Possible this is subjective, but I better like oldest type text spokening with this situation in heading level link. :-):-) This is a designed change related with heading level links, or this is an unwanted change? I using Ubuntu Lucid, Orca latest git master version. Attila =============== Created an attachment (id=167965) --> (https://bugzilla.gnome.org/attachment.cgi?id=167965) I temporary switch Orca morning state with not containing system voice feature This debug.out shows oldest speech generation with web browsing after I switch Orca with morning state. When I make this second debug.out file, Orca does'nt containing new system voice feature temporary, but shows very well the oldest speech generation with heading level links text. Attila
*** Bug 627042 has been marked as a duplicate of this bug. ***
So Attila, help me out here: Looking at the SPEECH OUTPUT lines, what is the difference? Is it a slight pause or something else? (i.e. is the wrong voice being used, in contrast to what the output suggests?)
Created attachment 167966 [details] [review] proposed fix for Paul's toggle button problem Solves it for me in OOo 3.3 dev. In theory it should solve it for 3.2 and earlier versions in which the toggle buttons are push buttons as far as we're told Anyhoo, please test. Thanks!
Created attachment 167968 [details] [review] possible fix for Paul's notification issue This solves it for me on my machine which has notify-osd installed. I have made a similar change for notification-daemon, which in theory should work, but it's not been tested. I'll need to find a machine on which to do so. Alternatively, if one of you happen to have notification-daemon (which is not notify-osd) installed and can test, it would be super cool. Thanks!
Joanie, Confirm the OO toggle patch works for version 3.2.0 of OO Thanks.
... and confirm the notification patch works. I'm also using notify-osd
Joanie, the problem is following: Oldest (before this afternoon) Orca releases sends heading level link text outputs with one Speech message. Now, this is changed following: Orca send for example following text if the caret positioned a heading level link: Joanie link Heading level 1 This is result two intonated sentence. Another problem (I found now when navigating Bugzilla page), I fraid this is true for more controls when the user browsing the internet, later need analyze and compare oldest and actual speech outputs: When the caret is positioned with a button in Firefox web browser, Orca send following type string: Commit button This is result broken intonation the "commit button" text spokening. Hearest the dot sentence with commit button control label (end of commit text), and hearest a dot sentence with button role type. This is not good. See for example following new speech generation string: SPEECH OUTPUT: 'Commit' SPEECH OUTPUT: 'gomb' SPEECH OUTPUT: 'Aktiválásához nyomja meg a szóköz billentyűt.' I think the problem is following this situation: Cutted two speech output part the commit and gomb (button role type hungarian translated message) text. If I remember right, before this feature Orca sends one speech message the commit button text, and sends another speech message for example the tutorial message if the Enable Tutorial check box is checked. Not possible the matter is the voice switch (sending default voice character the "commit" text, and after this Orca sending the role type with the system voice character? Sorry the long comment. :-):-) If need, I do tomorrow the comparisons with speech outputs with trivial form fields, if need I compare the oldest speech generation and new speech generation strings. Attila
Created attachment 167970 [details] This debug.out shows broken button intonation I found when producing Orca the broken form field control intonation: The problem is present only if I using the Orca jump prewious or next form field command (Orca+Tab or Orca+Shift+Tab key combination), or if I using up or down arrow key with webpage navigation. If I using Tab or Shift+tab key combination, I hear right intonation for example with button form fields. This debug.out demonstrate possible what happening if I navigating next or prewious form field in Bugzilla home page. Very interesting, Orca sometime two spokened the Find button text (find gomb in hungarian translation). Joanie, I hope this debug.out help found why happening the problem. Attila
It seems to me that the Welcome to Orca message is not spoken using the system voice, although the content of the file debug say the opposite. I think that orca is not using the correct parameters like rate and pitch. The Welcome to Orca is announced slowly.
Looking in the patch applied to notification-daemon-script.py seems to me that the following line should be included before 'speech.speak(utterances, None, True)' utterances.append(self.voices.get(settings.SYSTEM_VOICE)) Maybe I am wrong.
(In reply to comment #26) > Looking in the patch applied to notification-daemon-script.py seems to me that > the following line should be included before 'speech.speak(utterances, None, > True)' > > utterances.append(self.voices.get(settings.SYSTEM_VOICE)) > Maybe I am wrong. Yes and no. Appending the DEFAULT_VOICE explicitly would make sense and be more clear. If no voice is used, the DEFAULT_VOICE should be used, but I'm not convinced that is always occurring in Orca's existing speech and speechfactory modules. (I think that's the source of some of the weirdness people are reporting.) Anyhoo.... The reason it's not the system voice in that case is that we're getting labels and labels display actual text on the screen. Make sense? But I'll make the voice addition. Thanks for taking a look and catching that!!
Comment on attachment 167966 [details] [review] proposed fix for Paul's toggle button problem The toggle button fix for OOo has been committed to master. http://git.gnome.org/browse/orca/commit/?id=5177ab5f7e591d7633c41290ef5c5691a44ad431
Created attachment 167989 [details] [review] Just the notify-osd fix Based on Paul's feedback and my testing, it seems the fix for notify-osd is sound. Based on José's feedback, I need to add a voice to the notification-daemon script. So.... I have just committed to master the notify-osd fix. I will attach a new patch for notification-daemon. http://git.gnome.org/browse/orca/commit/?id=14310044f5a4d875c0ba8ca7119f2751e462c046
Created attachment 167990 [details] [review] Proposed patch - notification-daemon As I just commented, the notify-osd fix is in master. That leaves the fix for notification-daemon. This is the original proposed fix with the default voice explicitly appended to the utterances derived from the labels (i.e. displayed text). Thanks again José!
Created attachment 168001 [details] [review] Prefs reloaded fix and more * Says the prefs reloaded statement in the system voice * Says the welcome to orca message in the full system voice * possibly other fixes I mention the third item because the problem with the welcome to orca message is that it wasn't fully updating the voice. So I changed something small, but significant in the method that pretty much speaks everything. <smile> Please, please, please test. Thanks!
Created attachment 168003 [details] [review] Fixes the 'readonly' should be in the system voice issue This is very isolated and hence safe. And I've tested it quite a bit. Thus this one is already committed to master. http://git.gnome.org/browse/orca/commit/?id=01a70cb4334355db22de1b55ea406eed410ae2b8
Well I confirm the welcome and settings reloaded messages are now okay. If I notice anything else I'll comment back. Paul
(In reply to comment #24) > Very interesting, Orca sometime two > spokened the Find button text (find gomb in hungarian translation). Yeah, I was just going to say: Most of the time, the speech is collapsing things correctly -- even in your debug.out. And even when I arrow or use form field navigation that is the case as well. I've not yet tackled the general Gecko problem: We're speaking rolenames in web content in the incorrect voice. I plan to fix that soon (if possible). That may have additional impact on what you're reporting. But we'll figure it out. If you just cannot stand it in the meantime, perhaps using the stable version of Orca for web content. Just until we get this sorted out for you. > Joanie, I hope this debug.out help found why happening the problem. Nothing is jumping out at me yet (other than the fact that you're not always having the problem). Sorry.... But we will solve it with your help. I promise. Thanks!!
Comment on attachment 168001 [details] [review] Prefs reloaded fix and more Committed (modulo the extra whitespace). :-) http://git.gnome.org/browse/orca/commit/?id=0db7e4aba51ef3c9e7950bcd2b0536bbbcd28c0b
Before when I looked at the text widget 'blank' problem, I was seeing some serious weirdness. I think the previous commits (my guess is the most recent one), solved that. As a result, only a tiny change was needed. Orca master should now say 'blank' using the system voice. And there was much rejoicing. Yay. http://git.gnome.org/browse/orca/commit/?id=194b0472ee708fb2d07473ff5195cca9070cbcb8
(In reply to comment #36) > Before when I looked at the text widget 'blank' problem, I was seeing some > serious weirdness. I think the previous commits (my guess is the most recent > one), solved that. As a result, only a tiny change was needed. Orca master > should now say 'blank' using the system voice. And there was much rejoicing. > Yay. > > http://git.gnome.org/browse/orca/commit/?id=194b0472ee708fb2d07473ff5195cca9070cbcb8 When reading messages in thunderbird , blank lines are announced using the default voice.
After my last git pull, I am facing two problems: 1. When I press alt+tab, orca does not read the name of the next window. 2. The flat review mode no longer works. I found the following traceback in the debug output: SPEECH OUTPUT: 'Bug 543157 â It is not always possible to distinguish on-screen text from Orca's "system" messages - Namoroka' Traceback (most recent call last):
+ Trace 223263
s.processObjectEvent(event)
self.listeners[key](event)
self.presentStatusBar(event.source)
speech.speak(text)
_speak(content, acss, interrupt)
voice.update(acss)
Please pull master and try again. This is what I just announced on the list.
(In reply to comment #37) > (In reply to comment #36) > > Before when I looked at the text widget 'blank' problem, I was seeing some > > serious weirdness. I think the previous commits (my guess is the most recent > > one), solved that. As a result, only a tiny change was needed. Orca master > > should now say 'blank' using the system voice. And there was much rejoicing. > > Yay. > > > > http://git.gnome.org/browse/orca/commit/?id=194b0472ee708fb2d07473ff5195cca9070cbcb8 > When reading messages in thunderbird , blank lines are announced using the > default voice. Yup. Pretty much all things Gecko (including thunderbird) are still needing to be done. But thanks for the report on this issue!
(In reply to comment #39) > Please pull master and try again. This is what I just announced on the list. Works fine. I sent the comment before receive the message in the list. Thanks.
I thought an extra checkbox option was added to turn the system voice on or off was going to be added to the Orca UI. I don't remember if that was still in a patch or if that had been committed or if I just plain imagined it:). In any case, I did not find it while tabbing through the speech nor the general page. I would figure it would be added to the checkboxes before the synth selections.
Created attachment 168025 [details] This debug.out file demonstrate Orca+Tab, Orca+Shift+Tab wrong speech output generation problem, and demonstrate right working speech generation with Tab and Shift+Tab key Joanie, I maked a debug.out file with demonstrate the difference the Orca+Tab, Orca+Shift+Tab key generated wrong speech output schema, and shows right speech output generation with Tab and Shift+Tab key if I navigating form fields. I do following steps if I record events the debug.out file. 1. I launch Firefox, and open www.google.com webpage. 2. I jumps form fields with Orca+Tab key and Orca+Shift+Tab key with several times. When I do the jump form fields navigation, Orca sends following style the form fields with speech output, this is the wrong schema: SPEECH OUTPUT: 'Google keresés' SPEECH OUTPUT: 'gomb' SPEECH OUTPUT: 'Aktiválásához nyomja meg a szóköz billentyűt.' So, Orca splitted the spokened speech output with three unit if the caret positioned with a form field. 4. When this test step is completed, first I positioned the first form field, and press two time the Tab key. Now, Orca send right speech output schema, this is the right type speech output schema: SPEECH OUTPUT: 'Google keresés gomb.' SPEECH OUTPUT: 'Aktiválásához nyomja meg a szóköz billentyűt.' The difference the two pasted type speech output is following: Orca right doesn't split the control label and role type with two speech output unit, this is the right type spokened output. Help this comment you determining why happening this problem, or you need another information? I using latest updated Orca git master version. Attila
Joanie, I see another issue with producing similar wrong splitted three part speech output string: Please try jump links with following website: www.dumaalmas.hu When Orca says dumaalmas.hu link text, producing this type speech output: SPEECH OUTPUT: 'dumaalmas.hu' voice=hyperlink SPEECH OUTPUT: 'hivatkozás' Why not send Orca this speech output with one message? Another equals issue if a webpage containing an image hiperlink, reproduced with www.dumaalmas.hu webpage, see following wrong speech output: SPEECH OUTPUT: 'dumaalmas.hu' voice=hyperlink SPEECH OUTPUT: 'hivatkozás' SPEECH OUTPUT: 'kép.' voice=hyperlink Now I not want attach a debug.out file, because have attached prewious command a formfield problem related debug.out file. But, I copyed a part with this new recorded debug.out file, please look: ^^^^^ PROCESS OBJECT EVENT object:state-changed:active ^^^^^ DEQUEUED EVENT object:state-changed:focused <---------- vvvvv PROCESS OBJECT EVENT object:state-changed:focused vvvvv OBJECT EVENT: object:state-changed:focused detail=(1,0,None) app.name='Firefox' name='None' role='link' state='enabled focusable focused horizontal opaque sensitive showing visible visited' relations='' Finding top-level object for source.name= --> obj.name= --> obj.name= --> obj.name= --> obj.name=Dunaalmás hírporta | dumaalmas.hu --> obj.name= --> obj.name= --> obj.name= --> obj.name=Dunaalmás hírporta | dumaalmas.hu - Mozilla Firefox LOCUS OF FOCUS: app='Firefox' name='' role='link' event='object:state-changed:focused' KEYEVENT: type=0 id=65506 hw_code=62 modifiers=8192 event_string=(Shift_R) is_text=True timestamp=2353255 time=1282047474.110387 KEYBOARDEVENT: type=0 id=65506 hw_code=62 modifiers=0 event_string=(Shift_R) keyval_name=(Shift_R) is_text=True timestamp=2353255 time=1282047474.110891 orca.keyEcho: string to echo: Shift_R orca.isModifierKey: returning: True orca.isModifierKey: returning: True orca.isModifierKey: returning: True KEYEVENT: type=1 id=65506 hw_code=62 modifiers=8193 event_string=(Shift_R) is_text=True timestamp=2353349 time=1282047474.205232 KEYBOARDEVENT: type=1 id=65506 hw_code=62 modifiers=1 event_string=(Shift_R) keyval_name=(Shift_R) is_text=True timestamp=2353349 time=1282047474.205893 orca.isModifierKey: returning: True orca.isModifierKey: returning: True generate braille for unfocused app.name='Firefox' name='None' role='image' state='enabled horizontal opaque sensitive showing visible visited' relations='' (args={'formatType': 'unfocused', 'useDefaultFormatting': False, 'includeContext': False, 'role': ROLE_IMAGE, 'mode': 'braille', 'recursing': True}) using '(imageLink and [Link(obj, (asString(label + displayedText) or asString(name)) + " " + asString(value + roleName))] or [Component(obj, asString(label + displayedText + value + roleName))])' imageLink=<CORBA.Object 'IDL:Accessibility/Accessible:1.0' at 0x7fd858001b70> label=[] displayedText=[''] name=['dumaalmas.hu'] value=[''] roleName=['K\xc3\xa9p'] generate braille results: Link: 'dumaalmas.hu Kép', 0 BRAILLE LINE: 'dumaalmas.hu Kép h1' VISIBLE: '', cursor=0 generate speech for unfocused app.name='Firefox' name='None' role='link' state='enabled focusable focused horizontal opaque sensitive showing visible visited' relations='' (args={'formatType': 'unfocused', 'useDefaultFormatting': False, 'priorObj': <CORBA.Object 'IDL:Accessibility/Accessible:1.0' at 0x1d14a30>, 'role': ROLE_LINK, 'mode': 'speech', 'recursing': True}) using 'newAncestors + newRowHeader + newColumnHeader + newRadioButtonGroup + labelAndName + roleName + availability + newNodeLevel + unselectedCell + (tutorial and (pause + tutorial) or [])' script_utilities.commonAncestor... ...script_utilities.commonAncestor Object deemed to be useless: [section | ] Object deemed to be useless: [section | ] newAncestors=[] newRowHeader=[] newColumnHeader=[] newRadioButtonGroup=[] labelAndName=['dumaalmas.hu', {'average-pitch': 3.0}] roleName=['hivatkoz\xc3\xa1s', {'average-pitch': 3.0, 'rate': 80.0, 'gain': 10.0, 'family': {'locale': 'hu', 'name': 'hungarian'}}, 'k\xc3\xa9p', {'average-pitch': 3.0}] availability=[] newNodeLevel=[] unselectedCell=[] GENERATOR: getTutorial obj = role = link alreadyFocused = False utterances: () tutorial=[''] pause=[<orca.speech_generator.Pause instance at 0x17807e8>] generate speech results: dumaalmas.hu {'average-pitch': 3.0} hivatkozás {'average-pitch': 3.0, 'rate': 80.0, 'gain': 10.0, 'family': {'locale': 'hu', 'name': 'hungarian'}} kép {'average-pitch': 3.0} <orca.speech_generator.Pause instance at 0x17807e8> SPEECH OUTPUT: 'dumaalmas.hu' voice=hyperlink SPEECH OUTPUT: 'hivatkozás' SPEECH OUTPUT: 'kép.' voice=hyperlink Temporary how can possible disable the system log feature if I would like testing the oldest type speech generation with form fields and another web related controls (links, heading level links, image links, etc)? I think not good solution undo last commits local my machine. Not have a hidden toggle with disable this feature temporary? Attila
I noticed that when I do ctrl+alt+d for the desktop, Orca says "x-nautilus-desktop" in the default voice. This text isn't on screen. Also when I do alt+f1 for the main menu, I hear "top expanded edge panel" in the default voice. Actually, if I cycle between the panels with ctrl+alt+tab Orca always says the panel name in the default voice. And whenever I hear "icon view" such as when I open nautilus, it is in the default voice.
I have just created a new bug (bug 628837) for the Gecko system voice. As I stated there: ===== Given how late we are in the release cycle, I'm quite hesitant to make any changes which are not "high impact, low risk." Therefore, we will work on this issue in master after we branch for 2.32. Once the solution has been thoroughly tested, we can do a putback for 2.32.1. ===== I'll look at Paul's reported Nautilus issues.
(In reply to comment #45) > Actually, if I cycle between the panels with ctrl+alt+tab Orca always says the > panel name in the default voice. On this one, when you Ctrl+Alt+Tab, a window-like object appears and, depending on the window switcher, that actual text is displayed. Therefore I'm not sure what I can do about that issue. > And whenever I hear "icon view" such as when I open nautilus, it is in the > default voice. This issue should be fixed in master now. http://git.gnome.org/browse/orca/commit/?id=6137182677235ed996399b0d089f14c13b718571
(In reply to comment #45) > I noticed that when I do ctrl+alt+d for the desktop, Orca says > "x-nautilus-desktop" in the default voice. This text isn't on screen. The script in question is Nautilus. The accessible in question is of ROLE_FRAME. In Maverick, at least, the object's name is "Desktop" (and not x-nautilus-desktop). So from an event point of view, it's going to be challenging to distinguish: * Giving focus to the actual Desktop * Opening a window for the Desktop (via Places->Desktop) * Giving focus to the aforementioned window I can dig deeper, but.... Doing a bit of cost-benefit analysis, I'm not convinced it's worth the trouble (and likelihood that we guess incorrectly which situation we're in) to examine each and every frame in Nautilus for this one particular use case. If others disagree with me, however, then a new bug should be opened for this problem. > Also when I do alt+f1 for the main menu, I hear "top expanded edge panel" in > the default voice. I just opened bug 628842 for this issue.
I believe I have opened new bugs for all of the remaining issues. Therefore I'm closing this bug. Finally. If there are any additional issues, let's open new bugs for them as well. Thanks.
Readding blocks and dependencies for historical and research purposes.